This article presents an improved flow direction algorithm (FDA) called ROFDA, which integrates Lévy distribution and random opposition-based learning (ROBL) to enhance performance. The Lévy distribution refines the flow velocity vector, enabling more accurate movement toward optimal solutions through flexible and random velocity prediction. ROBL further strengthens the decision-making process by enhancing exploration and convergence. ROFDA is tested on 23 standard benchmark functions and applied to optimise the placement of electric vehicle fast charging stations (EVFCSs) and distributed generation (DG) units within 33- and 69-bus distribution systems (DS). Results show that ROFDA outperforms traditional FDA, OFDA, and MFDA methods. In systems with multiple EVFCSs of varying capacities and locations, it achieves over 70% reduction in power losses in some cases, while maintaining optimal voltage profiles. These improvements confirm ROFDA's effectiveness for ensuring reliable and efficient power delivery, making it a promising tool for EVFCS and DG planning.
Kumar et al. (Thu,) studied this question.